Author name cluster

David Watson

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

5 papers

1 author row

NeurIPS Conference 2025 Conference Paper

Abstract Counterfactuals for Language Model Agents

Edoardo Pona
Milad Kazemi
Yali Du
David Watson
Nicola Paoletti

Counterfactual inference is a powerful tool for analysing and evaluating autonomous agents, but its application to language model (LM) agents remains challenging. Existing work on counterfactuals in LMs has primarily focused on token-level counterfactuals, which are often inadequate for LM agents due to their open-ended action spaces. Unlike traditional agents with fixed, clearly defined action spaces, the actions of LM agents are often implicit in the strings they output, making their action spaces difficult to define and interpret. Furthermore, the meanings of individual tokens can shift depending on the context, adding complexity to token-level reasoning and sometimes leading to biased or meaningless counterfactuals. We introduce \emph{Abstract Counterfactuals}, a framework that emphasises high-level characteristics of actions and interactions within an environment, enabling counterfactual reasoning tailored to user-relevant features. Our experiments demonstrate that the approach produces consistent and meaningful counterfactuals while minimising the undesired side effects of token-level methods. We conduct experiments on text-based games and counterfactual text generation, while considering both token-level and latent-space interventions.

PDF Details

NeurIPS Conference 2025 Conference Paper

Autoencoding Random Forests

Binh Vu
Jan Kapar
Marvin Wright
David Watson

We propose a principled method for autoencoding with random forests. Our strategy builds on foundational results from nonparametric statistics and spectral graph theory to learn a low-dimensional embedding of the model that optimally represents relationships in the data. We provide exact and approximate solutions to the decoding problem via constrained optimization, split relabeling, and nearest neighbors regression. These methods effectively invert the compression pipeline, establishing a map from the embedding space back to the input space using splits learned by the ensemble's constituent trees. The resulting decoders are universally consistent under common regularity assumptions. The procedure works with supervised or unsupervised models, providing a window into conditional or joint distributions. We demonstrate various applications of this autoencoder, including powerful new tools for visualization, compression, clustering, and denoising. Experiments illustrate the ease and utility of our method in a wide range of settings, including tabular, image, and genomic data.

PDF Details

NeurIPS Conference 2023 Conference Paper

Explaining Predictive Uncertainty with Information Theoretic Shapley Values

David Watson
Joshua O'Hara
Niek Tax
Richard Mudd
Ido Guy

Researchers in explainable artificial intelligence have developed numerous methods for helping users understand the predictions of complex supervised learning models. By contrast, explaining the $\textit{uncertainty}$ of model outputs has received relatively little attention. We adapt the popular Shapley value framework to explain various types of predictive uncertainty, quantifying each feature's contribution to the conditional entropy of individual model outputs. We consider games with modified characteristic functions and find deep connections between the resulting Shapley values and fundamental quantities from information theory and conditional independence testing. We outline inference procedures for finite sample error rate control with provable guarantees, and implement efficient algorithms that perform well in a range of experiments on real and simulated data. Our method has applications to covariate shift detection, active learning, feature selection, and active feature-value acquisition.

PDF Details

NeurIPS Conference 2023 Conference Paper

Intervention Generalization: A View from Factor Graph Models

Gecia Bravo-Hermsdorff
David Watson
Jialin Yu
Jakob Zeitler
Ricardo Silva

One of the goals of causal inference is to generalize from past experiments and observational data to novel conditions. While it is in principle possible to eventually learn a mapping from a novel experimental condition to an outcome of interest, provided a sufficient variety of experiments is available in the training data, coping with a large combinatorial space of possible interventions is hard. Under a typical sparse experimental design, this mapping is ill-posed without relying on heavy regularization or prior distributions. Such assumptions may or may not be reliable, and can be hard to defend or test. In this paper, we take a close look at how to warrant a leap from past experiments to novel conditions based on minimal assumptions about the factorization of the distribution of the manipulated system, communicated in the well-understood language of factor graph models. A postulated interventional factor model (IFM) may not always be informative, but it conveniently abstracts away a need for explicitly modeling unmeasured confounding and feedback mechanisms, leading to directly testable claims. Given an IFM and datasets from a collection of experimental regimes, we derive conditions for identifiability of the expected outcomes of new regimes never observed in these training data. We implement our framework using several efficient algorithms, and apply them on a range of semi-synthetic experiments.

PDF Details

YNIMG Journal 2017 Journal Article

Fractionating the anterior temporal lobe: MVPA reveals differential responses to input and conceptual modality

Charlotte Murphy
Shirley-Ann Rueschemeyer
David Watson
Theodoros Karapanagiotidis
Jonathan Smallwood
Elizabeth Jefferies

Words activate cortical regions in accordance with their modality of presentation (i. e. , written vs. spoken), yet there is a long-standing debate about whether patterns of activity in any specific brain region capture modality-invariant conceptual information. Deficits in patients with semantic dementia highlight the anterior temporal lobe (ATL) as an amodal store of semantic knowledge but these studies do not permit precise localisation of this function. The current investigation used multiple imaging methods in healthy participants to examine functional dissociations within ATL. Multi-voxel pattern analysis identified spatially segregated regions: a response to input modality in anterior superior temporal gyrus (aSTG) and a response to meaning in more ventral anterior temporal lobe (vATL). This functional dissociation was supported by resting-state connectivity that found greater coupling for aSTG with primary auditory cortex and vATL with the default mode network. A meta-analytic decoding of these connectivity patterns implicated aSTG in processes closely tied to auditory processing (such as phonology and language) and vATL in meaning-based tasks (such as comprehension or social cognition). Thus we provide converging evidence for the segregation of meaning and input modality in the ATL.

Details DOI